U.S. patent number 4,129,856 [Application Number 05/742,048] was granted by the patent office on 1978-12-12 for filter system and method for intrusion alarm.
This patent grant is currently assigned to Contronic Controls Limited. Invention is credited to Peter E. Humphries.
United States Patent |
4,129,856 |
Humphries |
December 12, 1978 |
Filter system and method for intrusion alarm
Abstract
An intrusion alarm of the kind in which a radiation field is
transmitted in an area to be supervised; the reflected field is
received and compared with the transmitted field, and a comparison
signal is produced the frequency of which is indicative of the
speed of movement of an object in the supervised area. A signal
processor receives the comparison signal and generates an alarm if
the comparison signal contains components in a selected frequency
range. The processor includes a unique filter block which removes
unwanted frequency components outside the range. The filter block
includes a pulse generator which generates a constant amplitude
pulse train, each pulse corresponding to a pulsation of the
comparison signal, and control elements which control the duty
cycle of the pulse train so that as the frequency of the comparison
signal increases, the average level of the pulse train first
increases and then decreases. The pulse train is integrated and
operates an alarm signal generator when its average level rises
above a preselected level.
Inventors: |
Humphries; Peter E. (King,
CA) |
Assignee: |
Contronic Controls Limited
(Mississauga, CA)
|
Family
ID: |
24983302 |
Appl.
No.: |
05/742,048 |
Filed: |
November 15, 1976 |
Current U.S.
Class: |
340/554; 327/102;
342/28 |
Current CPC
Class: |
G01S
13/10 (20130101); G01S 15/523 (20130101) |
Current International
Class: |
G01S
15/52 (20060101); G01S 15/00 (20060101); G01S
13/00 (20060101); G01S 13/10 (20060101); G01S
009/02 (); G08B 013/16 () |
Field of
Search: |
;340/258R,258A,258B
;343/5PD,5DP ;307/233R,234 ;328/111,138,140,141 |
References Cited
[Referenced By]
U.S. Patent Documents
Primary Examiner: Caldwell, Sr.; John W.
Assistant Examiner: Nowicki; Joseph E.
Attorney, Agent or Firm: Rogers, Bereskin & Parr
Claims
What I claim is:
1. An intrusion alarm comprising:
(1) means for transmitting a radiation field in an area to be
supervised,
(2) means for receving a portion of the radiation field which is
reflected from objects in the supervised area,
(3) means for comparing the transmitted and received fields and for
producing a comparison signal having pulsations the frequency of
which is indicative of the speed of movement of an object in said
supervised area,
(4) means connected to said comparison means for generating an
output pulse train of constant peak amplitude, said output pulse
train having an average level, said means for generating including
means for controlling the duty cycle of said output pulse train for
said average level of said output pulse train to increase as the
frequency of said comparison signal increases, until the frequency
of said comparison signal reaches a selected frequency, and then to
decrease as the frequency of said comparison signal increases
beyond said selected frequency,
(5) integrating means, connected to said means for generating (4),
for integrating said output pulse train to produce an integrated
d.c. signal therefrom, said d.c. signal having an average
level,
(6) and means connected to said integrating means and responsive to
said average level of said integrated d.c. signal for producing an
alarm signal when such average level rises above a preselected
level.
2. An intrusion alarm according to claim 1 wherein said means for
generating includes means for producing pulses for said output
pulse train wherein each such pulse corresponds to a pulsation of
said comparison signal, and said means for controlling includes
means for controlling the duration of each said pulse of said
output pulse train so that the duration of such pulse is a
controlled duration equal to the lesser of (i) a fixed duration,
and (ii) the duration of its corresponding pulsation in said
comparison signal less a fixed interval.
3. An intrusion alarm according to claim 2 wherein said means for
controlling the duration of each said pulse of said output pulse
train includes:
(a) first generating means coupled to said comparison means and
operative on receipt of said comparison signal to generate a first
intermediate pulse train having constant maximum amplitude first
intermediate pulses, each corresponding to a pulsation of said
comparison signal, each first intermediate pulse having a leading
edge having a predetermined slope such that each intermediate pulse
requires said fixed interval to reach a predetermined level,
(b) second generating means connected to said first generating
means and operative during receipt of signals of level above said
predetermined level for thereby generating a second intermediate
pulse train having second constant amplitude intermediate pulses
each commencing at said fixed interval after the commencement of
its corresponding first intermediate pulse and each terminating on
termination of its corresponding first intermediate pulse,
(c) third generating means connected to said second generating
means and responsive to each second intermediate pulse for
generating a constant amplitude third pulse, each third pulse
commencing upon receipt of its associated second intermediate pulse
and terminating upon the ending of said controlled duration, said
third pulses together constituting said output pulse train.
4. An intrusion alarm according to claim 3 wherein said first
generating means includes a first amplifier having an output, a
resistance connected to said output, a capacitance connected to an
end of said resistance remote from said output terminal and adapted
to be connected to a power supply for said system, and a diode
connected in parallel with said resistance, said first amplifier
being connected to said comparison means (3) for producing constant
amplitude fourth pulses each corresponding to a pulsation of said
comparison signal, said diode being oriented so that it is reverse
biased upon its receipt of the leading edge of a said fourth pulse
and so that it is forward biased upon its receipt of the trailing
edge of such fourth pulse, so that said capacitance charges through
said resistor on receipt of the leading edge of a said fourth pulse
and discharges through said diode on receipt of the trailing edge
of such fourth pulse, and said second generating means comprising a
second amplifier having an input connected to said end of said
resistance and having an output, said second amplifier being biased
for conduction only when the voltage at its input rises above said
predetermined level, the output of said second amplifier being said
second intermediate pulse train.
5. An intrusion alarm according to claim 4 wherein said third
generating means comprises a third amplifier having an input, a
capacitance connected between said input of said third amplifier
and said output of said second amplifier, a resistance connected
between said input terminal of said third amplifier and adapted to
be connected to said power supply, and a diode connected in
parallel with said last mentioned resistance, said last mentioned
diode being oriented so that it is reverse biased when the leading
edge of a said second intermediate pulse is received at said input
of said third amplifier and so that it is forward biased when the
trailing edge of a said second intermediate pulse is received at
said input of said third amplifier, so that when the leading edge
of each said second pulse appears at the output of said second
amplifier, a corresponding step voltage is applied through said
last mentioned capacitance to the input of said third amplifier,
the voltage at the input of said third amplifier then varying as
said last mentioned capacitance charges through said last mentioned
resistance, and means biasing said third amplifier for conduction
only until the voltage at its input reaches a selected level, the
output of said third amplifier being said output pulse train.
6. An intrusion alarm according to claim 2 wherein said means for
transmitting transmits an ultrasonic sound field of frequency
greater than 30 KHz.
7. An intrusion alarm according to claim 2 wherein said means for
transmitting transmits an ultrasonic sound field of frequency
between 35 KHz and 45 KHz.
8. A method of detecting intruders comprising:
(1) transmitting a radiation field in an area to be supervised,
(2) receiving a portion of the radiation field which is reflected
from objects in the supervised area,
(3) comparing the transmitted and received fields and producing a
comparison signal having pulsations the frequency of which is
indicative of the speed of movement of an object in the supervised
area,
(4) generating an output pulse train of constant peak amplitude,
each pulse of which corresponds to a pulsation of said comparison
signal, said output pulse train having an average level,
(5) controlling the duty cycle of said output pulse train so that
the average level thereof increases as the frequency of said
comparison signal increases until the frequency of said comparison
signal reaches a selected frequency, and then decreases as the
frequency of said comparison signal increases beyond said selected
frequency,
(6) integrating said output pulse train to produce an integrated
d.c. signal having an average level,
(7) and generating an alarm signal when the average level of said
integrated d.c. signal rises above a preselected level.
9. A method according to claim 8 and including the step of
controlling the duration of each pulse of said output pulse train
so that its duration is equal to the lesser of a fixed duration and
the duration of its corresponding pulse in said comparison signal
less a fixed interval.
Description
This invention relates to apparatus and a method for detecting the
presence of an intruder and for producing an alarm signal when such
detection occurs.
Intrusion alarms commonly operate by transmitting an ultrasonic or
electromagnetic field, receiving a portion of the reflected field,
and comparing the two. If a moving intruder is present, a doppler
shift is created in the received field. The doppler shift is
detected and is used to produce an alarm indication.
One of the major difficulties with prior intrusion alarms has been
the frequency of false alarms. Disturbances such as rustling
curtains, telephone or other bells, the scraping of snow ploughs,
noisy motor bearings, or hot air from heating systems, all produce
disturbances which are likely to create a false alarm, particularly
when the transmitted field is ultrasonic sound. Active filters have
been used in an attempt to discriminate between such disturbances
and the disturbance caused by an intruder, but they have been
generally unsatisfactory. This is because the received field is
usually weak because of attenuation along its path of travel, so it
must be highly amplified before the disturbance signals can be
processed. If the amplitude of an unwanted disturbance signal is
sufficiently large, as it often is, the high amplification causes
part of the unwanted disturbance signal to leak through the filter,
triggering a false alarm.
The problem is made more acute when the transmitted field is
ultrasonic sound transmitted at higher frequencies. Higher
frequencies (typically above 30 KHz) are desirable since the
coverage of such fields is less affected by changes in the
temperature, density and humidity of the air through which they
travel. However, since the attenuation of higher frequency sound
waves is greater than that of lower frequency sound, even more
amplification is needed at higher frequencies and the likelihood of
false alarms increases correspondingly.
Accordingly, it is an object of the invention to provide an
intrusion alarm in which the likelihood of false alarms is reduced.
In a preferred embodiment of the invention this is achieved by
providing an intrusion alarm comprising:
(1) means for transmitting a radiation field in an area to be
supervised,
(2) means for receiving a portion of the radiation field which is
reflected from objects in the supervised area,
(3) means for comparing the transmitted and received fields and for
producing a comparison signal having pulsations the frequency of
which is indicative of the speed of movement of an object in said
supervised area,
(4) means connected to said comparison means for generating an
output pulse train of constant peak amplitude, and including means
for controlling the duty cycle of said output pulse train for the
average level of said output pulse train to increase as the
frequency of said comparison signal increases, until the frequency
of said comparison signal reaches a selected frequency, and then to
decrease as the frequency of said comparison signal increases
beyond said selected frequency,
(5) integrating means connected to said means (4) for integrating
said output pulse train to produce an integrated d.c. signal
therefrom,
(6) and means connected to said integrating means and responsive to
the average level of said integrated DC signal for producing an
alarm signal when such average level rises above a pre-selected
level.
Further aspects and advantages of the invention will appear from
the following disclosure, taken together with the accompanying
drawings, in which:
FIG. 1 is a block diagram of a system according to the
invention;
FIG. 2 is a frequency response curve for a portion of the system of
FIG. 1;
FIG. 3 is a schematic diagram of the system of FIG. 1;
FIGS. 4a to 4e show low frequency wave forms for the FIGS. 1 and 2
systems; and
FIGS. 5a to 5e show high frequency wave forms for the FIGS. 1 and 2
system.
GENERAL DESCRIPTION
The invention will be described with reference to an intrusion
detector which employs ultrasonic sound. Reference is first made to
the block diagram of FIG. 1, which shows a typical embodiment of
the invention. The FIG. 1 system includes an oscillator 2 which
provides the required ultrasonic frequency electrical signal and
which can be located with the remainder of the FIG. 1 system or
remotely in a separate master unit. The signal from the oscillator
2 may be of any appropriate frequency, typically 40 KHz.
The 40 KHz signal from oscillator 2 is used to drive a transmitter
4, which radiates a 40 KHz sound field into the area to be
supervised. The ultrasonic sound waves which are reflected from
objects and walls in the supervised area are received by a
transducer and receiver 6, the output of which is amplified by
amplifier 8.
The resultant amplified signal from amplifier 8 is compared with
the transmitted signal by a synchronous detector consisting of
transistor Q1. The base of transistor Q1 is connected to the
oscillator 2 so that transistor Q1 is switched on and off by the
same signal which drives the transmitter.
The signal at the collector of transistor Q1 is directed through a
band pass filter 10. The filter 10 attenuates frequencies below
about 40 Hz and also attenuates high frequencies, thus
substantially removing the 40 KHz component. Ideally the frequency
of the output signal from filter 10 now contains only a narrow band
of frequencies, namely those frequencies which are likely to be
generated by the movement of an intruder and not by other factors.
However, in practice the output signal from filter 10 usually also
contains other frequency components caused by unwanted disturbing
factors.
It is found that the doppler frequency caused by moving intruders
usually falls in the range 40 to 300 Hz when a 40 KHz sound field
is transmitted. It is found that under a wide range of conditions,
frequencies within the range 40 to 300 Hz in many unoccupied
premises being supervised are unlikely to be generated in the
comparison signal by anything but a moving intruder. Disturbances
at such premises, other than intruders, generally produce
components in the comparison signal which are outside that
frequency range.
Therefore, the signal from band pass filter 10 is amplified by
amplifier 12 and is then directed into a non-linear filter block
14. The non-linear filter block 14, which is a key part of the
invention and will be described in detail, produces a d.c. output
signal, the level of which depends on the frequency of the output
signal from the band pass filter 10. The frequency response of the
non-linear filter block 14 for a prototype of the present invention
is shown in FIG. 2, in which input frequency is plotted along the
horizontal axis (note that the scale shifts at 100 Hz to keep the
graph within manageable limits), and d.c. output voltage is plotted
along the vertical axis. The filter block 14 response 15 is
designed to rise sharply with input frequency to a peak at about 40
Hz, and then to fall off approximately linearly as the input
frequency increases beyond 40 Hz. Since the filter block 14 is
frequency dependant and is not affected by the amplitude of the
received signal (provided that the received signal is of adequate
amplitude), high frequency components in the input signal, even if
they are of high amplitude, do not produce an increase in
output.
The d.c. output from the filter block 14 is directed to a biased
voltage discriminator 16. The discriminator 16 is set so that it
produces an output only so long as its input voltage received from
the filter block 14 exceeds a predetermined level (typically level
18 in FIG. 2). Level 18 is selected to be the level reached by the
output of the filter block 14 when the filter block 14 receives an
input signal of frequency between about 40 and 300 Hz. Therefore,
the discriminator 16 will produce an output whenever an intruder,
for example, moves rapidly enough to produce a doppler shift of
between 40 and 300 KHz in the 40 KHz signal received by
transducer-receiver 6. The output from voltage discriminator 16
usually takes the form of one or more relatively short pulses, the
duration of which depend on the duration of movement of the
intruder. The pulses from discriminator 16 are therefore directed
to a pulse stretcher 20, which produces pulses of a predetermined
minimum length whenever it receives a pulse from discriminator 16.
The pulses from pulse stretcher 20 are applied to an alarm signal
generator 22 which produces an appropriate alarm.
DETAILED DESCRIPTION
Reference is next made to FIG. 3, which shows the circuit of FIG.
1, including the non-linear filter block 14, in more detail. In
FIG. 2, the values of typical components used are shown in
parenthesis beside the components.
As shown in FIG. 2, the signal from the collector of transistor Q1
is directed through resistor R1, capacitor C1 and resistor R2 to
the input 30 of amplifier 12, which is a high gain high impedance
operational amplifier. Capacitor C1 acts to attenuate very low
frequencies, while a capacitor C2 connected between resistor R1 and
ground removes high frequencies, including the 40 KHz driving
signal. Components R1, R2, C1, C2 together constitute the bandpass
filter 10.
The output from amplifier 12 is directed through capacitor C3 and
resistor R3 to input terminal 32 of a high gain amplifier 34.
Amplifier 34, which forms part of non-linear filter block 14, is
biased so that it saturates when it receives an input signal above
a minimum amplitude. The output signal from amplifier 34 is
therefore normally a square wave train, as shown at 36 in FIG. 4a.
Square wave train 36 swings between ground and +6 volts (the d.c.
supply level), and each positive pulse 38 of wave train 36
corresponds to the negative going half cycle of the input signal to
amplifier 34.
The output terminal 40 of amplifier 34 is connected to a non-linear
low pass filter 42 comprising diode D1, resistor R3 and capacitor
C3. The output of the low pass filter 42 is connected to the input
terminal 44 of a high gain amplifier 46. The operation of filter 42
is as follows. When wave train 38, FIG. 4a, goes positive as
indicated at 48, diode D1 is reverse biased, and capacitor C3
charges through resistor R4. Because of the time constant of the
charging circuit through resistor R4, the voltage at input terminal
44 of amplifier 46 rises relatively slowly. The slow rise is
indicated at 50 in FIG. 4b, which shows the voltage level at input
terminal 44 of amplifier 46. However, when the signal at the output
of amplifier 34 drops to zero, as indicated at 52 in FIG. 4a, diode
D1 of filter 42 becomes forward biased, permitting capacitor C3 to
discharge rapidly through diode D1, and the voltage level at input
terminal 44 of amplifier 46 falls rapidly. The rapid fall is
indicated at 54 in FIG. 4b. The result is that the pulses of the
wave train 55 at input terminal 44 of amplifier 46 have a
relatively slow rise time and a rapid fall time, as shown in FIG.
4b.
Amplifier 46 is biased so that it turns on only when the voltage at
its input terminal 44 exceeds a preset bias level indicated at 56
in FIG. 4b. The time interval 57 for the waveform 55 to rise to
this level is typically about 1 millisecond in the example
illustrated. Since amplifier 46 is an inverting amplifier, its
output at terminal 58 is shown as waveform 60 in FIG. 4c. As shown,
each pulse of waveform 60 begins later than its corresponding pulse
in waveform 36 by the duration of the fixed time interval 57.
The output terminal 58 of amplifier 46 is connected through a
capacitor C4 to input terminal 62 of an amplifier 64 which is
biased to conduct on negative input peaks only (as will be
explained). Input terminal 62 is also connected via a parallel
combination of diode D2 and resistor R4 to the positive voltage
supply.
The waveform at the input terminal 62 of amplifier 64 is shown at
66 in FIG. 4d. This wave form is produced as follows. When waveform
60 of FIG. 4c drops from plus 6 volts to zero, as indicated at 68,
the resultant negative going step is transmitted through capacitor
C4 to input terminal 62, reverse biasing diode D2. Capacitor C4
then charges slowly through resistor R4, causing a slow increase 70
(FIG. 4d) in the voltage at input terminal 62. When waveform 60
(FIG. 4c) goes positive again, as indicated at 72, a positive going
step is transmitted through capacitor C4 to terminal 62, driving
the voltage at this terminal positive as indicated at 74 in FIG.
4d. This forward biases diode D2, discharging capacitor C4 rapidly
so that the voltage at terminal 62 rapidly falls to 6 volts, where
it remains for the remainder of the half cycle.
Amplifier 64 is biased, as indicated by bias level voltage 76 in
FIG. 4d, so that it turns on only while the voltage at its input
terminal 62 is below the level 76. The resultant output wave form
at terminal 77 of amplifier 64 is shown at 78 in FIG. 4e. Waveform
78 consists of "on" intervals, or positive going square wave pulses
80, of approximately 9 milliseconds duration (the duration is
controlled by the discharge time of capacitor C4 through resistor
R4), separated by "off" intervals 82.
The waveform 78 at terminal 77 is directed through integrating
elements, namely resistor R5 and capacitor C5, to the input
terminal 84 of voltage discrimator 16. Resistor R5 and capacitor C5
integrate waveform 78, providing a relatively steady d.c. signal
the level of which is equal to the average level of waveform 78.
When the level of this d.c. signal exceeds the bias level 18 (FIGS.
3 and 4e) of voltage discriminator 16, the discriminator 16
produces an output.
It will be seen from inspection of the waveforms of FIG. 4 that the
"on" portions 80 of waveform 78 can never exceed the period timed
by capacitor C4 and resistor R4 (here about 9 milliseconds), and
the "off" portions 82 of waveform 78 will occupy the remaining part
of each cycle of the input waveform 36. When the frequency of
waveform 36 is low, the "duty cycle" or average level of output
waveform 78 will therefore be very low. As the input frequency
increases, the "duty cycle" of waveform 78, i.e. the ratio of the
duration of the "on" pulses 80 to that of the "off" pulses 82,
increases linearly, and therefore the average level of waveform 78
increases linearly with input frequency. This is shown as portion
90 of response curve 15, FIG. 2.
As the input frequency increases, the average level or duty cycle
of waveform 78 reaches a maximum when the duration of the maximum
"on" portion 80 of waveform 78, plus the fixed interval 57, is
equal to the duration of each "off" portion 82. The frequency at
which this occurs is marked at 92 in FIG. 2. Since the duration of
the maximum "on" portion 80 is 9 m.s. in the example illustrated,
and interval 57 is 1 m.s., frequency 92 occurs when each "off"
portion 82 is 9 + 1 = 10 m.s. in duration. This corresponds to an
input frequency cycle duration of 20 m.s., or 50 Hz.
As the frequency of input waveform 36 increases beyond 50 Hz,
waveform 66 (FIG. 4d) begins to go positive in each half cycle
before capacitor C4 charges to the bias level 76 (FIG. 4d).
Therefore, the duration of each "on" portion 80 of waveform 78
(FIG. 4e) is reduced. The duration of each "on" portion 80 now
becomes equal to the duration of each half cycle of input waveform
36 less the fixed interval 57. As the frequency of waveform 36
continues to increase, the fixed interval 57 becomes a larger
proportion of each "on" portion 80, and the duty cycle or average
level of output waveform 78 falls. This is shown as portion 94 of
frequency response curve 15, FIG. 2. (Portion 94 is in fact linear
but the horizontal scale of FIG. 2 changes at 100 Hz to keep the
length of the graph manageable.)
In general, it will be seen that the duration of each "on" portion
80 of waveform 78 is equal to either (a) 9 m.s. (determined by
capacitor C4 charging through resistor R4), or (b) the duration of
each half cycle of input waveform 36 less the fixed interval 57,
whichever is less.
By way of example, FIGS. 5a to 5e show waveforms corresponding to
those of FIGS. 4a to 4e (primed reference numerals are used to
indicate corresponding parts) but at a much higher input frequency.
As shown in FIG. 5b, the duration of each fixed interval 57 is a
substantial proportion of each half cycle 38 of the input waveform
36. The duty cycle, and therefore the average level, of waveform
78, has consequently become relatively low.
As indicated previously, the threshold at which discriminator 16
operates to produce an output is shown at 18 in FIG. 2. In the
example given, this corresponds to a range of frequencies of
between about 38 and 320 Hz for the input signal. Input signals
having a high frequency, no matter what their amplitude, will not
produce an output since the output is dependent entirely upon the
frequency of the input signal and not upon its amplitude (provided
that at least a minimum amplitude input signal is received). As a
result, the immunity of the system to spurious signals is much
improved. It is found that the system's immunity is for example
much improved for sounds produced by telephone and other bells,
which produce false alarms in most competitive systems.
When discriminator 16 produces an output, even of short duration,
this operates Schmidt trigger 100 in pulse stretch 20. Trigger 100
then produces a timed output pulse which constitutes an alarm
signal and may be used as desired. As shown in FIG. 1, the pulse
may be used to operate an alarm signal generator 22.
An additional advantage of the system described is that it enables
ultrasonic operation at higher frequencies. Most ultrasonic alarm
systems operate at a broadcast sound field frequency of between 20
KHz and 26.5 KHz. Higher frequencies are generally avoided because
they are attenuated to a greater extent in air. The higher
attenuation requires greater amplification, resulting in a greater
likelihood of false alarms. However, apart from the false alarm
difficulty, higher frequencies (30 KHz and up) are desirable
because they are less affected by changes in atmospheric conditions
such as air pressure, humidity and termperature. For example, if an
intrusion alarm is adjusted under conditions of low humidity and
high air pressure to cover a specified area, then under conditions
of high temperature and humidity with resultant low air pressure,
the pattern of protection may be only 70% of the original setting.
Conversely, if a system is adjusted on a summer day to avoid
sources of interference in the room, such as air conditioners, then
on a winter day, the pattern of protection will extend into areas
which were not intended to be covered, and the likelihood of false
alarms increases. Therefore, to avoid resetting the system every
time atmospheric conditions change, it is desirable to broadcast at
increased frequencies, which are less subject to such changes. The
unique filter block of the invention enables operation at
frequencies higher than 25 KHz. For example, operation may be at 30
KHz or more, with much reduced likelihood of false alarms. An
operating frequency of between 35 and 45 KHz is preferred, and 40
KHz has been selected as a particularly suitable frequency. Of
course the filter block can also be used with lower frequency
systems.
Although the filter block of the invention has been described with
reference to an ultrasonic intrusion alarm, it is also applicable
to an intrusion alarm using electromagnetic radiation. Such
intrusion alarms operate in the same manner as ultrasonic intrusion
detectors, and again the filter block of the invention may be used
to eliminate undesired frequency components from the comparison
signal.
It will be appreciated that although the system described operates
when the average level of the d.c. signal from filter block 14
rises above level 18 in FIG. 4e, the system can be inverted so that
the average level of the d.c. output from filter block 14 is a
maximum at frequencies outside the range of interest and is below a
predetermined level in the frequency range of interest. In the
appended claims, therefore, the term "increase" as used with
reference to signal levels, includes an increase in a positive
sense and an increase in a negative sense, as may be
appropriate.
* * * * *